ABSTRACT
Background: The survival of patients with stage IA-IIA non-small cell lung cancer (NSCLC) after surgery is heterogeneous. This study aimed to construct a prognostic risk model to predict the overall survival (OS) of these patients. Methods: Data from patients (n=9,914) from the Surveillance Epidemiology and End Results (SEER) database were analyzed. The cases were randomly divided into the training and the validation groups. Patients from the Shanghai Pulmonary Hospital (n=270) were also included as an external cohort. Independent significant factors affecting survival in the training cohort were used to construct a nomogram. The precision was evaluated using the concordance index (C-index) and calibration plots. The X-tile software was used to confirm the optimal cut-off value to classify the patients. Results: Sex, age at diagnosis, tumor size, visceral pleura invasion (VPI), tumor grade, and the number of examined lymph nodes were deemed independent prognostic factors and were selected to establish the nomogram. The C-indices of the nomogram for predicting OS were 0.671 [95% confidence interval (CI): 0.653-0.689] in the training group, and 0.668 (95% CI: 0.650-0.687) and 0.707 (95% CI: 0.651-0.763) in the validation and the testing groups, respectively. The cut-off value of risk points was 106.0, which stratified the patients into high-risk and low-risk groups. The high-risk patients had shorter 5-year OS than low-risk patients (P<0.001). Conclusions: The established nomogram could evaluate the survival in patients with stage IA-IIA NSCLC after surgery and may provide prognostic information for clinicians to make decisions in the management of adjuvant therapy.
ABSTRACT
This study constructed and validated a prognostic model to evaluate the survival of small-cell lung cancer (SCLC) patients following surgery, and shed light on the strategy of postoperative radiotherapy. A total of 882 patients from Shanghai Pulmonary Hospital and the Surveillance, Epidemiology and End Results database after lung resection were selected. Multivariable Cox analysis was used to identify the indicators affecting long-term survival in patients. A nomogram was constructed to predict the prognosis of eligible patients. Indices of concordance (C-index) was used to access the predictive ability of cancer-specific survival (CSS) for the prognostic model. CSS discrimination in the prognostic model was comparable in the training and validation cohorts (C-index = 0.637[NORAD-T], 0.660[NORAD-V], 0.656[RAD] and 0.627[our hospital], respectively. Stratification based on the cutoff value of the nomogram yielded low- and high-risk subgroups in four cohorts. For patients in the high-risk group, postoperative radiotherapy was considered a survival-promoting strategy (unadjusted HR 0.641, 95% CI 0.469-0.876, p = 0.0046). In the low-risk group, however, the implementation of radiotherapy barely had an influence on CSS. In conclusion, the nomogram we constructed and validated could predict the prognosis of SCLC patients followed surgery and identify high-risk patients who were likely to benefit from postoperative radiotherapy.
